Changes in the ocular and nasal signs and symptoms of aircrews in relation to the ban on smoking on intercontinental flights

Changes in the ocular and nasal signs and symptoms of aircrews in relation to the ban on smoking on intercontinental flights. Scand Objectives This study determined the influence of exposure to environmental tobacco smoke (ETS) in aircraft on measured and perceived cabin air quality (CAQ), symptoms, tear-film stability, nasal patency, and bioinarkers in nasal lavage fluid. Methods Commercial aircrews underwent a standardized examination, including acoustic rhinometry, nasal lavage, and measurement of tear-film break-up time. Eosinophilic cationic protein, myeloperoxidase, lysozyme, and albumin were analyzed in the nasal lavage fluid. Inflight investigations [participation rate 98% (N=39)] were performed on board 4 flights, 2 in each direction between Scandinavia and Japan. Scandinavian crew on 6 flights from Scandinavia to Japan participated in postflight measurements after landing [participation rate 85% (N=41)]. Half the flights permitted smoking on board, and the other half, 0.5 months later, did not. Hygienic measurements showed low relative air humidity on board (2-10%) and a carbon dioxide concentration of <I000 ppm during 99.6% of the cruising time. Results The smoking ban caused a drastic reduction of respirable particles, from a mean of 66 (SD 56) pg/m3 to 3 (SD 0.8) yg/m3. The perceived CAQ was improved, and there were fewer symptoms, particularly ocular symptoms, headache and tiredness. Tear-film stability increased, and nasal patency was altered. separation nonsmokers, as indicated by a large increase in respirable particles. This ETS exposure is associated with an increase in ocular and general symptoms, decreased tear-film stability, and alterations of nasal patency.

Previous questionnaire investigations of commercial aircrews (19)(20)(21)(22) have found a high prevalence of eye, nose, and throat symptoms, typically 40-70%. In 2 of the studies (21,22), the symptoms were suspected to be caused by exposure to ozone during higher altitude flights. In addition, there has been concern about environmental tobacco smoke as a health hazard in aircraft (10,(12)(13)(14)(15)(16)(17)(18). Recently, many flight companies have prohibited smoking on board, but few hygienic or medical evaluations have been made of the ban on smoking in aircraft.
Recently, objective methods have been developed with which to study environmental effects on the upper airways and eyes (23)(24)(25). Acoustic rhinometry can be applied in field studies of nasal patency (26,27). In addition nasal lavage is a well documented technique for studying inflammatory effects in the nasal mucosa in relation to inhalatory exposure (24)(25)(26)(27). Experimental studies with nasal lavage have shown that a large number of possible biomarkers is available, including tryptase, albumin, lysozyme, eosinophil cationic protein emitted from the granula of activated eosinophilic granulocytes, and myeloperoxidase emitted from activated neutrophilic granulocytes (24,25,28). In addition, measurements of objective signs from the eyes (eg, tear-film stability) can be applied (29-3 1). Recent epidemiologic studies have applied these objective methods (26,27,29,31), but none have been used in studies on aircrews.
In this study our aim was to determine the influence of a ban on smoking on commercial airlines on measured and perceived cabin air quality, nonspecific symptomatology, tear-film stability, nasal patency, and selected biomarkers in nasal lavage fluid.

Subjects of the inflight investigation
The research team made 2 trips direct from Stockholm to Tokyo, and back, either via Copenhagen or directly to Stockholm. One trip was made at the end of August, just before inflight smoking was banned; the others took place in mid-September, 2 weeks after smoking was banned on all intercontinental flights. On each flight, there are always 2 Japanese attendants, 5 Scandinavian cabin attendants, and 3 Scandinavian pilots. All aircrew members on these 4 flights (N=40) were invited to participate in the medical investigations performed on board during the later part of the 11-to 12-hour flights, operated by the Scandinavian Airline System (SAS). Altogether, 39 out of 40 (98%) answered the medical ques-tionnaire, and 37 (93%) participated in the medical examinations. One Japanese crew member with asthma was excluded on a smoking flight when the examination was restricted to nonasthmatic subjects.

Subjects of the postflight investigation
To increase the study population, and to enable a comparison of the differences in clinical parameters after landing, postflight aircrews of 6 flights staying at the same crew hotel in Tokyo were examined postflight. Two of the flights had the same staff as in the inflight investigation, the other 4 were flights arriving on the same days as the flight used by the research team. The medical examinations were performed at the same crew hotel in central Tokyo, on the 12th floor of the hotel, mostly within 24 hours after the flight. Since the Japanese crew from each flight did not stay at the crew hotel, they were excluded from this part of the study. Altogether 41 of the 48 Scandinavian aircrew members (85%) participated in the postflight investigation. One Scandinavian crew member with asthma, who worked on a smoking flight and participated only in the postflight investigations, was excluded when the examination was restricted to nonasthmatic subjects.

Assessment of personal factors and symptoms
The subjects were questioned by a physician about allergy and other diseases, medication, occupational data, smoking habits, and social status. Atopy was defined as a current history of allergic manifestations related to exposure to common allergens mediated by immunoglobulin (Ig) E in Sweden (tree pollen, grass pollen, or animal fur) reported in the medical interview. Information on current symptoms during flights was gathered by means of a self-administered questionnaire, given to the subjects by the physician at the time of the medical examination. The set of symptoms included questions on nasal symptoms, ocular symptoms, throat symptoms, dermal symptoms, dyspnea, and general symptoms. There was also a question about dyspnea during the flight. To avoid differences in pollen exposure in Scandinavia between smoking and nonsmoking flights, the examinations were performed at the end of August and in mid-September, an off-pollen season.
Altogether, 5 1 aircrew members participated in the study, either in the inflight or the postflight investigation. Of these persons, one-third consisted of pilots. All the pilots were men, about one-fourth of all the participants were current smokers, and about 1 out of 7 had a history of atopy. Frequent sinusitis was common, reported by 25% of all the participants. None of the participants used contact lenses during flights. One crew member participating in the inflight investigation only and another participating in the postflight investigation only were asthmatics who had been diagnosed by a physician. Both were on smoking flights, none from the nonsmoking flights had asthma. The personal characteristics of the crews on smoking flights were similar to those of the nonsmoking flight crews (

Aircraft characterization
The same type of major intercontinental aircraft (Boeing 767-300) with a total number of 190 seats was used on all the flights. The smoking seats in the tourist class (rows 21-39) were located near the aft galley, and smokers in the Euroclass (rows 1-17) were located near the middle section. The aircraft had a cabin volume of 428 m3, and a calculated ventilation capacity of 1320 llsecond (32). The ventilation system normally provided approximately 50% fresh air and 50% recirculated air to the passenger cabin. Normally, the airflow rate should have provided about 7 turnovers per hour of fresh air to the cabin, or about 10 liters of outdoor air per passenger per second from the outside and the same flow of recirculated air in the passenger cabin. Occasionally, the pilot can shut off the recirculation fans for short periods, the result being a 100% fresh air supply and a 2-fold increase in the outdoor airflow. The air exchange rate on the flight deck is about 60 turnovers per hour. The engines in the flight supply the fresh air, passing through heated zones in the engine (400°C). The air is conditioned by the air-conditioning packs, and ozone is removed as the fresh air passes through a catalytic ozone converter. The cabin air circulation system utilizes a prefilter for larger particles, a high-efficiency particulate air filter (HEPA) capturing particles equal or greater than 0.3 ym with 99.99% efficiency, and a charcoal filter to remove volatile compounds. The air-conditioning system does not contain any air humidification devices.

Methods to measure climate and air pollution in the cabin and at the crew hotel
Inflight measurements were performed simultaneously with the medical examinations on board. The measurements included temperature, relative air humidity, concentration of respirable particles in the air, carbon dioxide (CO,), airborne molds and bacteria, ozone, and formaldehyde. Volatile organic compounds (VOC) were measured, but the results will be reported in a separate publication. In addition, room temperature, relative air humidity, respirable particles, airborne molds and bacteria, ozone, and formaldehyde were measured in the room in which the postflight medical examinations took place at the crew hotel.
Inflight temperature and air humidity were recorded with a SWEMA logger 15 (SWEMA AB, Sweden), which sampled l-minute average values. The logger was calibrated at the factory in connection with the investigation. Concentrations of respirable particles were measured by a direct-reading instrument, based on light scattering (Sibata P-5H2, Sibata Scientific Technology Ltd, Japan), that had been used in earlier indoor air investigations (33). The CO, concentration was measured by a direct-reading infrared spectrometer (Rieken RI-41 lA, Rieken Keini, Japan), calibrated by standard gases containing known concentrations of CO,. The signal from the dust monitor and the CO, meter were also recorded with the SWEMA logger, which sampled 1-minute average values. Moreover, short-term measurements of temperature and air humidity were done manually by an Assman psychrometer, and 15-minute average values of respirable particles were recorded during the medical investigation.
Airborne microorganisms were sampled on 25-mm nucleopore filters with a pore size of 0.4 ym and a sampling rate of 1.5 llminute for 4 hours. The total concentrations of airborne molds and bacteria were determined by the CAMNEA method (34). Viable molds and bacteria were determined by incubation on 2 different media. The detection limit for viable organisms was 30 colonyforming units (cfu) per cubic meter of air. Ozone was measured with another diffusion sampler from the Swedish Environmental Research Institute, Sweden (35). To achieve sufficient detection limits, the sampling time was expanded to 16 hours, by sampling during 2 flights on the same sampler. The concentration of formaldehyde was measured with glass-fiber filters impregnated with 2,4-dinitro-phenylhydrazine (36), the air sampling rate being 0.2 llminute for 4 hours. The filters were analyzed by liquid chromatography.

Indoor climate and air pollution in the aircraft
When the medical examinations were made, the mean room temperature in the forward galley was 23.1°C dur- ing the smoking flights and 23.0°C during the nonsmok- Table 3. Respirable pal-kicle concentrations before and after the ing flights. In the aft galley, the room temperature was ban on smoking. room was 12 pg/m3 for respirable particles, 5 pg/m3 for formaldehyde, and 5.5 yglm" for ozone. No detectable air concentrations of viable molds were found in the hotel (< 100 cfulm"), the concentration of viable bacteria was low (1 10 cfu/m3), and there was no detectable total mold concentration (<I0 000 organisms/m3) and a low concentration of total bactelia (12 000 organisms/m3). No viable species of molds or bacteria could be identified.  (table 3).
The average ozone concentration during all flights I Acousfic rhinomefry was 13.5 yg/m3 in the aft galley, 32.5 pg/m3 in the forward galley, and 60.9 pg/m3 in the cockpit. The concentration of formaldehyde, in both the forward and aft gallies was below the detection limit (<5 yg/m3) during all the flights, irrespective of the smoke or nonsmoke conditions. The concentrations of viable molds and bacteria, as well as the total mold and bacteria concentrations, were measured on 1 flight after the smoking ban. In the aft galley, neither air concentrations of viable molds or bacteria (< 100 cfn/m3) nor total concentrations of molds or bacteria (<I0 000 organisms/m3) could be detected.
In the forward galley, there were no detectable air concentrations of viable molds (< 100 cfu/m"), a low concentration of viable bacteria (140 cfu/m3), and no detectable total mold or bacteria concentration (<I0 000 organisms/m3) was found. No viable species could be identified for any sample during the cultivation.

Indoor climate and air pollution at the crew hotel
The average room temperature in the examination room was 23.4"C in August 1997 after the smoking flights and 23.9"C in September 1997 after the nonsmoking flights, the difference being nonsignificant. The average relative air humidity in the hotel room was 65% in both August and September. The concentration in the air of the hotel Acoustic rhinometry (Rhin 2000, SR Electronics, Denmark; wideband noise; continuously transmitted) was performed at the end of the flight. The measurements were made according to standardized forms (sitting), after 5 minutes of rest, and prior to the lavage (26,27). By means of acoustic reflection the minimal cross-sectional areas (MCA) on each side of the nose were measured from 0 to 22 mm (MCA1) and from 23 to 54 min (MCA2) from the nasal opening. The volumes of the nasal cavity on the right and left sides were also measured from 0 to 22 mm (Voll) and from 23 to 54 mm (Vo12). The mean values were calculated from 3 subsequent measurements on each side of the nose, and the data on nasal dimensions are presented as the sum of the values from the sight and left sides.

Nasal lavage
Lavage of the nasal mucosa was made with a 20-ml plastic syringe attached to a nose olive. The subjects were standing, with their head flexed about 30 degrees forward (26,27). Room-temperature (20-22"C), sterile 0.9% saline solution was introduced into the nasal cavity. Each nostril was lavaged with a 5-ml solution that was flushed back and forth 5 times via the syringe, at an interval of a few seconds. The fluid was transferred to a 10-ml polypropylene centrifuge tube. It was kept on ice, and within 300 minutes the solution was centrifuged at 800 g for 5 minutes. The supernatant was recentrifuged at 1400 g for 5 minutes and immediately frozen to -20°C. Lysozyme was analyzed by means of radioimmunoassay (38). Eosinophil cationic protein and myeloperoxidase were measured by means of a double antibody radioimmunoassay method (Pharmacia Diagnostics AB, Uppsala, Sweden) (39,40). The intra-and interassay variation coefficients for all 3 tests were less than 11%. Albumin was measured by rate nephelometry on an array protein system (Beckman Instruments Inc).

Tear-film stability
Tear-film break-up time was estimated by a standardized method, self-reported break-up time, or break-up times measuring the time the subjects could keep their eyes  open without pain when watching a fixed point on the wall. The method has been used previously in field studies (27,31), and its results have been shown to correlate well with those of the fluoresceine method for detecting tear-film break-up time (31,41).

Statistical analyses
Differences in the dichotomous personal characteristics and symptoms between the aircrews of the smoking and nonsmoking flights were analyzed by Fishers' exact test. The differences in age, total symptom index, and all types of clinical signs were tested by the Mann-Whitney Utest. The differences in the average values from the 1minute measurements of the cabin climate and the respirable dust concentrations between the smoking and nonsmoking flights were analyzed by student's t-test. In all the statistical analyses, 2-tailed tests and a 5% level of significance were used.

Symptoms and subjective cabin air quality in relation to inflight exposure to environmental tobacco smoke
In all the statistical analyses, the 2 subjects with asthma from the smoking flights were excluded. A numerical decrease of all types of individual symptoms, except facial dermal symptoms, occurred for the nonasthmatic subjects after the smoking ban (table 4). When the material was dichotomized with respect to the occurrence of more than 1 symptom in each symptom category, the occurrence of more than 1 ocular symptom was decreased from 55% to 11% (P=0.004) after the smoking ban (table 5). In addition, the total symptom score was higher for the smoke conditions (mean 3.9, SD 4.0) than for the nonsmoke conditions (mean 1.4, SD 1.6) (P=0.05). The symptom improvement was the most pronounced for the general symptoms. Before the smoking ban, 35% of the crew reported at least 1 general symptom (headache, fatigue, nausea, sensation of catching a cold). After the ban, none reported any of these symptoms (P=0.005 by Fisher's exact test). In addition, the perception of the cabin air quality was improved after the ban (P=0.03), while the ratings for perceived air humidity and dustiness remained unchanged (table 6).

Physiological signs in relation to exposure to environmental tobacco smoke
The measurement of tear-film stability during the flights showed a significantly increased stability after the smoking ban (P=0.01), with a mean of 25 seconds as compared with 10 seconds. There were also alterations in nasal patency in relation to the smoking ban. Rhinometsic measurements on board showed a greater posterior nasal cross-sectional area (MCA2) during the smoking flights than during the nonsmoke conditions (P=0.05), but no significant differences for the other nasal dimensions (MCAI, Voll, Vo12) were found (table 7).
Postflight measurements in the hotel showed a numerical but insignificant increase in tear-film stability after the smoking ban (a mean of 42 seconds as compared with 25 seconds). The rhinometric measurements at the hotel showed a numerically greater nasal patency for all 4 parameters after the nonsmoking flights, statistically significant for anterior nasal volume (Voll) (P=0.03) (table 7). Nasal lavage, performed at the crew hotel, showed no significant effects of smoking on board. The median concentration of eosinophil cationic protein was similar during both conditions [median 1.0 (interquartile range <I-1.

Discussion
Our study showed that exposure to environmental tobacco smoke in commercial aircraft may impair cabin air quality and increase exposure to respirable particles with an order of magnitude of > l . We also showed that exposure to environmental tobacco smoke affects aircrews, is the first publication on the hygienic and physiological effects of the ban on smoking in commercial aircraft.
The study was interventional, but, since the aircrew changes from flight to flight, none of the participants appeared more than once in the investigation. Selection bias due to low response rate was less likely since the participation rate was relatively high, 85% in the postflight clinical examination and 98% in the study on board. Moreover, the participants on board during the smoking and nonsmoking conditions were comparable with respect to mean age, gender, tobacco smoking, atopy, and occupation. The aircraft were the same, seasonal variation between smoking flights and nonsmoking flights should be negligible, the medical investigations were standardized, and all samples from the nasal lavage were run in the same batch. It could be possible that smokers would avoid flying on nonsmoking flights, due to withdrawal symptoms. However, in our study, no significant difference in the proportion of smokers was observed between the smoking and nonsmoking flights. It could also be possible that sensitive staff, not able to stand exposure to environmental tobacco smoke during smoking flights, would start flying on intercontinental flights after the ban on smoking. We had 2 asthmatics in the crew during the smoking flights and none during the  nonsmoking flights. Since the numerical differences in asthmatics between the groups could have biased the results of a small study, further statistical analysis was restricted to nonasthmatic aircrew members. Recall bias due to an awareness of exposure may affect symptom reporting, but it is unlikely to affect physiological signs. Thus we do not believe that our conclusions have been seriously biased by selection or response errors, or due to chance findings. In addition, the possibility that the observed relationships could be explained by some factor other than the exposure to environmental tobacco smoke on board the flights is less likely. We found that irritative symptoms of different types were common among aircrews, but less common in the absence of environmental tobacco smoke. Even a higher symptom prevalence has been reported in earlier studies on aircraft crew. These studies were performed during previous decades, when ozone exposure could have been a major problem. We carried out a study on an aircraft (Boeing 767-300) with ozone converters, and we found low mean ozone levels in the cabin. In a North American survey of 744 flight attendants, 95% reported eye discomfort in aircraft (19). In another American study on 1330 flight attendants, 60% reported eye problems, 75% reported sore or dry throat, and 16% reported nose bleeding (21). A high prevalence of nasal and ocular symptoms (43-74%) was also reported in another American study (22). In a subsequent study on 1240 Norwegian aircrew members employed by the Scandinavian Airlines System (SAS), 53% reported eye irritation, 44% complained of nasal congestion, and 26% had sinusitis (20).
Despite the efficient air filtration system, there was a striking effect of smoking on the concentration of respirable dust. We found an average concentration of 66 pg/m3 of respirable particles, with maximum values up to 253 yg/m" in the aft galley during smoke conditions. The smoking ban reduced the mean concentration to 4 yg/mqn the forward galley and to 3 yg/m3 in aft galley. The respirable particle concentrations in the aircraft after the ban were lower than in the hotel room in Tokyo (12 yg/m3) and lower than the mean concentration in Swedish dwellings (19 yg/m3), measured by the same instrumentation (42). Smokers and nonsmokers are separated in both Euroclass and tourist class. Since smokers in tourist class are seated next to the aft galley, environmental tobacco smoke is obviously transported into the aft galley. We have recently showed significant exposure to environmental tobacco smoke, through the detection of cotinine in urine, among aircrews working in the aft galley, while those working in the forward galley had no significant decrease of cotinine in urine after intercontinental flights in a Boeing 767-300 (15). The results of earlier studies, mainly focusing on passengers' exposure to environmental tobacco smoke, have resembled the results of our study, smoking flights having 5-to 20-fold higher levels of respirable suspendable particles at passengers' seats during smoking flights than during nonsmoking flights (12,14,17).
We found that environmental tobacco smoke in aircraft decreases the preceived cabin air quality and increases the occurrence of symptoms, particularly eye symptoms, headache, and tiredness. There was a significantly increased tear-film stability after the cessation of smoking, from a mean value of 10 seconds to 25 seconds. Despite the low relative air humidity on board, tearfilm stability after the smoking ban was numerically higher among the flight crew than among office workers (median 17 seconds), when measured by the same method (27), and it partly normalized after landing.
The rhinometric data suggested alterations of nasal patency in relation to exposure to environmental tobacco smoke on board, with an acute decongestive effect of such smoke in aircraft, followed by a congestive rebound effect after landing. The explanation for this finding is unclear, but it could be a combined effect of environmental tobacco smoke and lower air humidity, temperature, and air pressure in the aircraft as compared with the hotel conditions. In other studies performed during normal pressure and higher relative air humidity ( 4 0 4 2 % ) an acute congestive effect of environmental tobacco smoke was observed (6,7).
Nasal lavage, performed only in the postflight investigation at the crew hotel, showed no significant effects of smoking on board. For practical reasons, the nasal lavage was not performed directly after landing, and therefore transient nasal effects could have been missed. Median lavage fluid concentrations of biomarkers in the aircrew members, irrespective of the presence or absence of environmental tobacco smoke, were only 30-70% of the mean values reported for 83 office workers in mid-Sweden. This study used the same nasal lavage technique, and the same clinical laboratory, as for our aircrew investigation. For office workers, the median concentrations of eosinophil cationic protein, myeloperoxidase, lysozyme, and albumin were 1.4 pg/l, 5.5 pg/l, 0.70 mg/l, and <2 mg/l, respectively (27).
The relative air humidity was very low in both the aft and forward galleys (2-lo%), much lower than the relative air humidity indoors (33-75%) in mid-Swedish dwellings during the heating season (43). Other studies have also found low relative air humidity in aircraft, typically 5 4 9 % (10, l l , 4 4 ) , but not as low as our values, measured during cruising at a high altitude (33-37 000 feet; 10-1 1 2776 meters), near the North Pole with extremely low outdoor humidity.
In addition to the high levels of respirable particles during smoke conditions and the low relative air humidity, other exposure measurements showed low or normal values. The concentration of molds and bacteria was low both in the aircraft and at the hotel, less than 10% of av-erage levels of molds and bacteria measured in Swedish dwellings, measured by the same method (4.3). Other investigators, using different methods to measure microbial contamination, have reported similar findings (12,45). We found that the CO, concentration on board, an indicator of the outdoor air supply rate, was generally between 500 and 700 ppm. Only 0.3% of the measured 1minute values were above the current Swedish ventilation standard of 1000 ppm (37). The low concentration of CO, and microorganisms illustrates that the air filtration and ventilation systems had sufficient efficiency in the particular type of aircraft used. Other investigators have reported higher values, with average concentrations of 785-1756 ppm of CO, and 25-87% of all measurements above 1000 ppm (32).
In conclusion, tobacco smoking in commercial aircraft causes a drastic increase in the exposure to respirable particles. This exposure to environmental tobacco smoke may cause irritative symptoms, particularly ocular and general symptoms, impair the perception of cabin air quality, decrease tear-film stability, and alter nasal patency in aircrew members. Our results support the view that, despite the high air exchange rate and efficient air filtration, smoking in commercial aircraft leads to significant pollution in aircraft. It also implies that the measurement of tear-film stability and the use of acoustic rhinometry, in combination with the determination of specific biomarkers in nasal lavage fluid, can be applied in field studies on the medical effects of the cabin environment of aircraft.
This study was partly supported by grants from the Swedish Council for Worklife Research and The Swedish Foundation for Health Care Sciences and Allergy Research.